ata: do some threading-related corrections (some sync related where it matters). Make sure 'sleeping' is 'false' before the powerup sequence (throws-off powermgmt and it *is* about to woken again and powermgmt will need to compensate). Avoid looking at mutex flag directly; there's no tangible benefit I can discern and changes to the kernel that alter the meaning shouldn't be able to break anything. For that, in the jz sd driver, have sd_disk_is_active just return 'false' like all the other SD drivers do. If it *must* return this, it should implement it's own method but it seems to not matter.

git-svn-id: svn://svn.rockbox.org/rockbox/trunk@26234 a1c6a512-1295-4272-9138-f99709370657
This commit is contained in:
Michael Sevakis 2010-05-21 15:33:31 +00:00
parent e24dd5ff98
commit ff25b3a208
2 changed files with 23 additions and 22 deletions

View file

@ -1395,7 +1395,7 @@ void sd_sleepnow(void)
bool sd_disk_is_active(void)
{
return sd_mtx.locked;
return false;
}
int sd_soft_reset(void)